Chameleon Chips

You’ve seen the commercials. Soon you’ll be doing business on the beach, complete with audioconferencing and instantaneous access to the latest sales data from the office. What this vision of a wireless future disregards, though, is that high-end portable devices have a voracious need for power because each application requires its own silicon chip. Maybe wireless computing has a future, but it’s only going to last as long as your batteries do.

Silicon Valley-based QuickSilver Technology, one of many startups working on the problem, has found a way to drastically cut power requirements by making a single silicon chip that can reconfigure itself on the fly to understand different wireless signals. Developed in partnership with BellSouth and scheduled for midyear release, QuickSilver’s first chip will be able to decode incompatible cell-phone standards so that a user can access different phone networks instead of just one. Eventually the company expects to have chips that can run multiple applications, switching from audio to video to text processing-all on the juice of a cell phone. “What we’re really doing is building a container,” says Paul Master, vice president of technology at QuickSilver. “The value is the software you pour into it.”

The QuickSilver chip includes a programmable logic “fabric,” an array of logic gates that can understand software signals and reprogram itself for different operations. When a chip receives, say, a specific cellular phone signal, it calls instructions to support that signal’s protocol from its memory, loads the appropriate architecture into the reconfigurable portion of the chip and executes the operation. When it receives a different type of signal, it reconfigures itself once more and repeats the process. By comparison, a device that used a different piece of silicon for each operation would drain a battery in no time-that’s why it’s currently difficult to play MP3 files on your cell phone.

If QuickSilver does deliver, then yet another important piece of the wireless computing puzzle will have fallen into place. “They’re trying to do something that hasn’t been done before, and it’s not clear they’ll succeed,” says Nick Tredennick, editor of Dynamic Silicon, a newsletter that covers dynamic logic devices. “But of course, you could say that about all great ideas.”